EVAPORATION AND CONDENSATION 297 



Here M is the molecular weight of the vapor, R is the gas constant, and 

 r is the coefficient of reflection of the vapor molecules which strike the 

 surface of tlie solid (or liquid). It was stated at that time (1913) that 

 "There are good reasons for believing that the reflection of vapor molecules 

 from the surface takes place to a negligible degree only," and there followed 

 a discussion of Knudsen's results on the heat conductivity of gases at low 

 pressures which we shall consider in more detail below. 



The above formula, taking r = o, was then used to calculate the vapor 

 •pressure of tungsten from the experimental data on the rate of evaporation. 



Very recently, however, Knudsen ^ has accurately determined the rate 

 of evaporation of mercury in a high vacuum, and by comparing this with 

 the vapor pressure, using an equation similar to ( i ) , he has been able to 

 prove conclusively that, with a clean mercury surface, r does not exceed 

 one per cent. That is, he finds that all the mercury atoms striking a mercury 

 surface are condensed when the surface is at room temperature. 



This result of Knudsen's, together with Wood's conclusion that re- 

 flection is absent when the glass plate is at the temperature of liquid air, 

 indicates that there is little or no reflectivity of mercury atoms from either 

 a solid or liquid surface of mercury. 



Experiments on the formation of "shadows" in the evaporation of 

 filaments in a high vacuum have proved that at least for surfaces at room 

 temperature the same conclusion may be drawn regarding the reflectivity 

 of atoms of tungsten, platinum, copper, gold, silver, molybdenum, carbon, 

 iron, nickel and thorium. 



The reflectivity of metal atoms striking surfaces of the same metals 

 at room temperature (or lower) is zero. 



The question remains whether the reflectivity is still zero when the 

 surfaces are at higher temperatures and when surfaces other than that of 

 the metal are concerned. Knudsen's work with mercury has shown that 

 the reflectivity with this metal remains zero even at temperatures where 

 the vapor pressure of the metal becomes fairly high. 



We shall see, however, from a consideration of other cases, that the 

 coefficient of reflectivity changes very little if at all with temperature, so 

 that I think the above rule may safely be generalized to include metal 

 vapors condensing on metals at any temperature. 



Whether or not the same rule would apply to non-metallic substances 

 with equal accuracy, is rather doubtful. However, we shall see from 

 theoretical considerations that even in such cases as that of hydrogen atoms 

 striking a surface of liquid or solid hydrogen, there is reason to believe 

 that the amount of reflection is small. 



■* Ann. Phys., 47, 697, IQ15. 



